“The other way it manifests doesnt really eat up much ram, its more like a fragmented hard disk, as time goes on if a piece of code doesnt clean up properly after itself, what its needs gets splattered around like clothes in a teenage boys bedroom, this instead manifests as a slowly decreasing frame rate, until after a few hours of simming you finally hit just a few fps and rage quit. Default 744 in XP11 has this (is fixed in the sparky744).”

My understanding is that Windows 11 is constantly defragging the hard drives and that no exterior program is needed to do that. Am I wrong or do I misunderstand?

Sherm

31 minutes ago, shermank said:

My understanding is that Windows 11 is constantly defragging the hard drives and that no exterior program is needed to do that. Am I wrong or do I misunderstand?

Same problem, different part of the system.

Here Im talking about RAM, the chips that store data that are very, very quick to read and write data to, but lose everything when you restart.

Xplane install on the hard disk is say 100GB,

With 8GB of ram, obviously not all of that can be loaded at the same time, so what happens is data is continually being loaded and unloaded from the hdd into ram (and vram) as it is needed.

This is where the "blurries" come from, not enough ram/vram to load the full high resolution image, so it loads it into the system as low resolution that needs less space.

That process creates fragmentaion the same way writing and deleting files to a hard disk creates fragmentation, but organising it is (almost) entriely the job of the piece of code that is juggling the data.

Interesting responses and issues. It seems that after a few years of XP12 what was a good system with lots of memory buffers has been eroded away. Depends on who you read as to the culprit there seems to be a view that LR need to do more the way XP uses RAM and multicore chips. The again the developers merrily coding away - which is what I came up against - need more RAM go get it and it will work. Checked as much as I can and as I can get very good performance with my most complex models (multi-engine and lots of screen displays all synchronised) like bad weather, complex scenery and model still gives ,me 50-60FPS with Xp12. So bit more checking showed no serious bottle necks between CPU and GPU but yes the RAM side of things could be a bit stressed (now and then never longer than for about 1 sec) So closer look showed that to run a single engine piston with no advanced electronics with particular developer meant doubling my RAM but continuing on with a GPU and aged NVIDIA card that struggles not at all and performs beautifully with a quarter of the memory of the CPU.

Yep all those beautiful textures and ways of setting them up comes with quite a memory penalty so I guess the point was do I really want to go and spend a heap of cash on upgrading a system that is running beautifully as I see it to merely load the best and greatest . MY point was that the RAM issue had become so severe with the developer that the model does not load, floods RAM and hangs the PC. (Not CTD). Apparently with hundreds of happy customers I was the only malcontent complaining (Main question was why does the upgrade bust a model I have used for years and not the others?) Note not seeking refunds or anything else just to get the thing working was all!

As I said hardly impressed because the developer should make clear that what they are doing is likely to need a major system upgrade to work for me not let me go through countless hours and back and forth with them as well to be told - cannot fix it your systems is stuffed get a new one. Then I looked at 16GB of RAM on the CPU and 8GB on the GPU and thought so if the main CPU is that bad how come not the GPU?

It just means really that this developer has lost a good customer for the time being. Maybe in a year or two I may upgrade but in the meantime I will live with what I have got - works fine for all the other stuff!!

It turns out my problems were caused byas is almost always the case—new Nvidia drivers and an error with Vulkan.

1 hour ago, coastaldriver said:

Then I looked at 16GB of RAM

Just pick up a couple of 16gb or better 32gb sticks, despite the media hype they only cost like €20 to €40 each and are very easy to swap.

1 hour ago, coastaldriver said:

Interesting responses and issues. It seems that after a few years of XP12 what was a good system with lots of memory buffers has been eroded away.

16 GB RAM was never enough for X-Plane, at least as soon as scenery add-ons like orthos, X-World and HD Mesh entered the picture.

1 hour ago, coastaldriver said:

MY point was that the RAM issue had become so severe with the developer that the model does not load, floods RAM and hangs the PC. (Not CTD).

Any good Linux distro auto-terminates applications if they demand a lot more memory than there is available. Hanging the entire system points to something else, either a hardware fault or opeerating system issue.

1 hour ago, Aglos77 said:

It turns out my problems were caused byas is almost always the case—new Nvidia drivers and an error with Vulkan.

It's as if nvidia is vibe coding their drivers these days. Seems like it's a perpetual mess now.

This is something Coastaldriver should try out. Per his own description, it's obviously not RAM exhaustion and simply loading up CPU usage doesn't lead to a full system hang like he's describing. He's blaming a 3rd party dev for his own system issues. Unfortunate.

Most mainstream distros provide ability to revert nvidia driver version. They don't usually have that capability for Mesa because Mesa is quick to fix any bugs. And somehow AMD does a much better job of maintaining their linux kernel-space drivers than they do in windows, or than nvidia does in general right now.

On 8/13/2026 at 4:15 PM, coastaldriver said:

if the Felix B747-200 (4 engine jet) can function normally and without problems then why is this basic piston engine GA model such a problem

Are we talking about Thranda…?

The thing with some devs is that they’re just not interested in edge cases - esp if it happens to highlight their sloppy code.

I re 16gb, yeah, I feel you. Truth is I won’t even let my clients buy Windows laptops with less than 32gb these days(!). Tho in fairness, a lot of that is due to Googles monstrous Chrome.

In any case, my sympathies for running into that brick wall. Always frustrating, especially with something that’s supposed to bring you joy. 🙏

15 hours ago, blingthinger said:

He's blaming a 3rd party dev for his own system issues. Unfortunate.

As described, the issue only occurs with this particular addon. All others run acceptably.

I’d say blaming the dev is entirely fair game.

On 8/14/2026 at 2:19 PM, mldavis2 said:

Those who choose to run Linux for flight sims have always had issues. There are entire forums dedicated to trying to get things to work. LR does not program X-Plane for Linux, it is built on Macs and ported to Windows. Everything on Linux is a work-around, and trying to run on 16GB of RAM is very risky. Does the developer not publish recommended system specs for their products?

I run 64GB of RAM on Windows exactly for those reasons.

As someone who is using Fedora Linux for quite some time, I can not confirm this to be the case. X-Plane 12 runs perfectly well on my system and its Radeon RX 7900 XT. X-Plane has a native Linux binary, so no workaround like Proton or Wine needed. I also have 64 gigs, and I have no issues with any plane, library, or scenery.

This is not to evangalize a Linux system or something - I am just saying that for me, on my system, what you describe has never been the case.

8 hours ago, UrgentSiesta said:

I’d say blaming the dev is entirely fair game.

Not yet. There's no firm evidence. Circumstantial? Yes, but that's not enough to convict. Addon acft don't take down an entire OS unless there's some odd library issue or maybe something in the kernel. Maybe in that sense, it's a simple as starting the sim with the command line arg --zink to force on the zink layer. None of that is necessarily a dev's fault.

Thus far we've heard that ram and cpu usage are possibly the culprits but no actual data to implicate either (rant != data). 100% usage of either of those doesn't take down an OS unless there's hardware failures or overclocking in the mix.

Easy to over diagnose the problems - still it is only me I guess. Checked my physical hardware - no issues. RAM limited - maybe. Not up todate GPU driver from NVIDIA?

Well the outcome is simple - running XP12 on my 16GB RAM and 8GB GPU in either full screen or windowed modes - with all sliders for graphics at max except shadows results in the sim running at 50 FPS no stutters, no antialiasing visuals, clean clocking and time keeping. So I actually do not have a sim or PC issue. Leave it at that for the time being.

As for the developer - their choice to load up the sim models with junk!

7 hours ago, blingthinger said:

Not yet. There's no firm evidence. Circumstantial? Yes, but that's not enough to convict. Addon acft don't take down an entire OS unless there's some odd library issue or maybe something in the kernel. Maybe in that sense, it's a simple as starting the sim with the command line arg --zink to force on the zink layer. None of that is necessarily a dev's fault.

Thus far we've heard that ram and cpu usage are possibly the culprits but no actual data to implicate either (rant != data). 100% usage of either of those doesn't take down an OS unless there's hardware failures or overclocking in the mix.

Too focused on the tree and not enough on the forest.

If everything works fine until that particular addon is loaded, then…

Hey if that's how you folks need to label the scientific method, so be it.

I have also spent enough time in and diagnosticating the land of linux to know that that explanation isn't passing the sniff test.

my rant is lOOKING LIKE A dead horse now!